The impact of expected variance in performance on the satisfaction process

Research in economics, finance and decision science has shown that consumers are familiar with unit‐to‐unit variability, and in the context of services it has been demonstrated that consumers often anticipate and perceive performance heterogeneity. However, satisfaction models to date have failed to explicitly treat expectations as distributions. In this study, expectations were modeled along two dimensions – mean and variance of expected performance – which were manipulated together with actual performance in a true experimental design. The findings indicate that the expected variance in performance had an impact on perceived disconfirmation. Specifically, at low levels of incongruity (i.e. small absolute performance deviations from the expected mean), a high expected variance in performance reduced the level of perceived disconfirmation. Conversely, at high levels of incongruity (large absolute performance deviations from expectations), the expected variance in performance exerted minimal influence over perceived disconfirmation. These findings are reconciled and discussed using the zones of indifference and tolerance, and assimilation processes.